Security audit

Toeic Practice

Security checks for vulnerabilities and agentic risk

Overview

This is a narrowly scoped TOEIC practice skill with minimal state tracking and no executable code or network behavior.

Before installing, understand that the skill will keep a small local state file to rotate daily TOEIC Parts 5, 6, and 7. It is suitable for TOEIC reading practice, but users wanting non-English explanations may need to ask for them explicitly.

A.I.G

Vulnerability Patterns

Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns

Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (5)

Vague Triggers

Medium

Confidence: 91% confidence
Finding: The example trigger phrases are broad enough to match ordinary study-assistance requests, which can cause the skill to activate in situations beyond the user's explicit intent. Over-broad invocation increases the chance of prompt/skill routing mistakes and can override more appropriate skills or default safety behavior, though the educational TOEIC context limits the direct harm.

Vague Triggers

Medium

Confidence: 88% confidence
Finding: The manifest description says to use the skill when Hana should 'deliver TOEIC homework, drill Parts 5 to 7, or explain why an answer is correct,' but it does not define explicit trigger phrases, scope limits, or exclusion conditions. Phrases like 'explain why an answer is correct' are broad and could overlap with many ordinary tutoring interactions beyond this specific daily TOEIC workflow.

Natural-Language Policy Violations

Medium

Confidence: 96% confidence
Finding: The instruction 'Keep all practice content in English' imposes a fixed language choice without offering the user a language or explanation-language preference. This can violate language/locale policy when the skill does not provide opt-in or an alternative for explanations in the user's preferred language.

Natural-Language Policy Violations

Low

Confidence: 78% confidence
Finding: The README is organized under an "English" section and describes the skill as an English TOEIC coach, but it does not explicitly state whether explanations or interaction language can be adapted to the user's preference. This can conflict with language/locale policy expectations when no user opt-in or choice is documented.

Missing User Warnings

Low

Confidence: 90% confidence
Finding: The skill instructs the agent to overwrite a persistent state file on each run without any user-facing notice or consent. While the stored content is minimal, silent persistence can violate user expectations and becomes more concerning if the same storage area is reused by other skills or if users are unaware that interaction state is being retained.

Static analysis

No suspicious patterns detected.